Bone and fat tissue in children and adolescents:

Bone and fat tissue in children and adolescents:

studies with focus on osteocalcin

Bojan Tubić

Department of Pediatrics Institute of Clinical Sciences

Sahlgrenska Academy at University of Gothenburg

Gothenburg 2016

nervosa by Bojan Tubić

Main supervisor:

Diana Swolin-Eide, Associate Professor, University of Gothenburg Co-supervisors:

Per Magnusson, Professor, Linköping University

Staffan Mårild, Associate Professor, University of Gothenburg Göran Wennergren, Professor, University of Gothenburg

Bone and fat tissue in children and adolescents: studies with focus on osteocalcin

© Bojan Tubić 2016 bojan.tubic@gu.se

ISSN 978-91-628-9878-6 (printed)

ISBN 978-91-628-9879-3 (epub)

http://hdl.handle.net/2077/43460

Printed in Gothenburg, Sweden 2016

Ineko AB

To my family

“The scientific man does not aim at an immediate result.

He does not expect that his advanced ideas will be readily taken up.

His work is like that of a planter – for the future.

His duty is to lay the foundation for those who are to come, and point the way.

He lives and labors and hopes. ”

– Nikola Tesla, 1900

The general aim was to investigate the possible interplay between bone and fat tissue through clinical studies of children and adolescents. Osteocalcin (OC), a bone formation marker, has been proposed to act as a link between bone and energy metabolism in mice, but human data are inconclusive. The specific aims of this thesis were: (i) to clarify the role of OC in relation to weight, with focus on undercarboxylated OC (ucOC) and carboxylated OC (cOC); (ii) to gain insight on how obesity and underweight affect bone and fat tissue in children and adolescents and; (iii) to study the effect of whole body vibration (WBV) on parameters of metabolic syndrome, bone metabolism and body composition in children with obesity. Methodology:

Children and adolescents aged 2-24 years were included in the four studies.

Study I and II were cross-sectional (case-control), and study III and IV were interventional with a 12-week follow-up, of which study IV was a randomized case-control study. Biochemical parameters were examined in all four studies. Bone mass and body composition were assessed by dual-energy X-ray absorptiometry (DXA), peripheral quantitative computed tomography, heel DXA and laser. Methods of intervention were high-energy diet in patients with anorexia nervosa (AN) and WBV in patients with obesity.

Results: Total OC and ucOC did not differ between normal-weight and overweight subjects; however, overweight subjects had lower cOC levels, and the measured OC forms did not correlate with insulin and glucose.

Overweight children had increased bone mineral content (BMC) and bone mineral density (BMD) in comparison with normal-weight children, and there was a positive correlation between BMC, BMD and body mass index standard deviation score. Adiponectin was inversely correlated with BMC and BMD, and was an independent determinant of BMC and BMD. Patients with AN gained in weight and levels of all three forms of OC and BMC increased. The WBV did not result in any anthropometric changes; however, a reduction of sclerostin implies that WBV therapy has direct effects on bone

mechanotransduction.

Conclusions: This thesis could not confirm the hypothesis that OC has a positive effect on glucose and insulin homeostasis, although cOC was lower in obese subjects than in normal-weight subjects. The home-based WBV intervention study in young children with obesity did not result in any effect on weight, metabolic parameters or calcaneal bone mass.

Keywords: Osteocalcin, obesity, adiponectin, carboxylation, anorexia nervosa, paediatric, bone turnover markers, bone mass, whole body vibration,

muscle

ISBN: 978-91-628-9878-6 (printed)

Ökad inaktivitet och felaktig kost orsakar livstilsrelaterad ohälsa som fetma och insulinresistens. Fetma är bl a associerat med diabetes, hjärt- och kärlsjukdomar. I kontrast till fetma finns anorexia nervosa (AN) med dess allvarliga somatiska och psykiska följder. Syftet med avhandlingen var att undersöka det endokrina samspelet mellan skelett- och fettvävnad genom kliniska studier på barn och ungdomar. Målsättningen var att studera hela spannet av vikttillstånd, från fetma till AN. Experimentella djurstudier och kliniska studier har visat att benformationsmarkören osteocalcin (OC) har en positiv effekt på metabola parametrar som t ex insulin och glukosnivåer.

Fokus var att klargöra den föreslagna positiva rollen för OC med inriktning på de olika OC formerna, underkarboxylerat OC (ucOC) och karboxylerat OC (cOC). Vidare undersöktes hur fetma och undervikt påverkar skelett- och fettvävnaden hos unga och slutligen studera vibrationsträning och dess inverkan på muskler, ben och glukosmetabolism på barn med fetma.

I delstudie I deltog 28 normalviktiga och 13 överviktiga/obesa barn där vikt, längd, hälbenets bentäthet samt blodprover undersöktes. Bentätheten var högre hos överviktiga jämfört med normalviktiga barn och positivt korrelerat med standardiserat “body mass index”. Adiponektin visade negativt samband med bentäthet men det var inte någon skillnad i grupperna avseende OC. I delstudie II inkluderades barn (2-9 år), 62 överviktiga/obesa och 46 normalviktiga kontroller, ur den svenska delen av IDEFICS studien. Totalt OC och ucOC visade ingen skillnad mellan grupperna. Karboxylerat OC var lägre hos överviktiga. Totalt OC och cOC var negativt korrelerat till HbA1C.

I delstudie III studerades 22 patienter med svår AN under 12 veckors intensiv viktuppgångsbehandling. Vikt, längd, blodprover samt bentäthetsmätningar undersöktes vid studiestart och -slut. Totalt OC, ucOC samt cOC ökade men ingen av OC formerna korrelerade med viktförändring eller insulinnivåer. Mängden benmineral ökade. I delstudie IV deltog 30 överviktiga barn, 7-17 år (16 kontroller), i en prospektiv, randomiserad, kontrollstudie där ena gruppen genomgick vibrationsträning under 12 veckor.

Vikt, längd och blodprover insamlades. Bentätheten i hälen, muskelstyrka och balans mättes. Vikten och bentätheten var oförändrad men balansen förbättrades.

Sammanfattningsvis har dessa studier inte kunnat konfirmera en positiv

metabola roll för OC vid fetma och insulinresistens. Noterbart är att

karboxylerat OC var lägre hos överviktiga.

Povećana neaktivnost i nepravilna ishrana dovodi do zdravstvenih problema uzrokovanih faktorima životnog stila kao što su gojaznost i rezistencija na insulin. Gojaznost uzrokuje dijabetes i kardiovaskularne bolesti. Suprotno od gojaznosti je anoreksia nervoza (AN) sa ozbiljnim somatskim i psihološkim posledicama. Cilj ovog naucnog rada sprovodenog kliničkim studijama nad decom i adolescentima je bio da se ispita endokrinološka interakcija između koštanog i masnog tkiva. Namera je bila da se prouče dve grupe pacienata, gojaznih i anorektični. Eksperimentalna istraživanja nad životinjama i kliničke studije su pokazale da osteokalcin (OC), marker formiranja kostiju, ima pozitivan efekat na metaboličke parametre poput nivoa insulina i sećera u krvi. Ideja je bila da se potvrdi predpostavljeni pozitivni uticaj OC sa naglaskom na različite oblike OC (nedovljno karboksiliran OC (ucOC) i karboksiliran OC (cOC)). Potom je ispitivano kako gojaznost i pothranjenost utiču na kosti i masno tkivo kod mladih kao i uticaj treninga sa vibracijama na mišiće, kosti i metabolizam sećera kod gojazne dece. U studiji I je uključeno 28 dece sa normalnom težinom i 13 gojazne dece sa merenim parametrima: težina, visina, gustina petne kosti i uzorak krvi. Gustina kosti bila je veća kod gojaznih u poređenju sa decom normalne težine i pokazala je pozitivnu korelaciju sa standardizovanim "indeksom telesne mase".

Adiponektin je pokazao negativnu korelaciju sa koštanom gustinom. OC se nije razlikovo u grupama. U studiji II uključeno je 62 gojazne i 46 dece sa normalnom težinom kao kontrolna grupa (uzrast 2-9 godina stari), izabrana iz IDEFICS studije u Švedskoj. Nivo ukupnog OC i ucOC se razlikovo između grupa. Nivo karboksiliranog OC je bio niži kod gojaznih. Ukupni OC i cOC je pokazao negativnu korelaciju sa HbA1c. U studiji III ispitivalo se 22 bolesnika sa teškom AN tokom 12 nedelja intenzivnog režima ishrane za povećane telesne težine. Težina, visina, uzorak krvi i merenje gustine kostiju ispitivani su na početku i kraju studije. Nivo ukupnog OC, ucOC i cOC je bio povećan, ali nijedan od oblika OC nije pokazao korelaciju sa promenom težine ili nivoom insulina. Nivo mineralnog koštanog sadržaja je bio povišen.

U studiji IV ispitivano je 30 gojazne dece (uzrast 7-17 godina), od kojih su

14 bila kontrolna grupa, u prospektivnoj, randomiziranoj, kontroliranoj

studiji. Grupa sa16 dece je bila podvrgnuta treningu sa vibracijama u toku 12

nedelja. Mereni parametri su: težina, visina i uzorci krvi, gustina petne kosti,

mišićna snaga i ravnoteža. Gustina petne kosti i težina je bila nepromenjena,

dok je ravnoteža bila poboljšana. Kao zaključak, ove studije nisu mogle da

potvrde pozitivni metabolički efekat OC na gojaznost i insulinsku

rezistenciju. Zapažen je niži nivo osteokalcina kod gojazne dece.

Повећана неактивност и неправилна исхрана доводи до здравствених проблема узрокованих факторима животног стила као што су гојазност и резистенција на инсулин.

Гојазност узрокује дијабетес и кардиоваскуларне болести. Супротно од гојазности је анорексиа нервоза (АН) са озбиљним соматским и психолошким последицама. Циљ овог науцног рада спроводеног клиничким студијама над децом и адолесцентима је био да се испита ендокринолошка интеракција између коштаног и масног ткива. Намера је била да се проуче две групе пациената, гојазних и аноректични. Експериментална истраживања над животињама и клиничке студије су показале да остеокалцин (ОЦ), маркер формирања костију, има позитиван ефекат на метаболичке параметре попут нивоа инсулина и сећера у крви. Идеја је била да се потврди предпостављени позитивни утицај ОЦ са нагласком на различите облике ОЦ (недовљно карбоксилиран ОЦ (уцОЦ) и карбоксилиран ОЦ (цОЦ)).

Потом је испитивано како гојазност и потхрањеност утичу на кости и масно ткиво код младих као и утицај тренинга са вибрацијама на мишиће, кости и метаболизам сећера код гојазне деце. У студији I је укључено 28 деце са нормалном тежином и 13 гојазне деце са мереним параметрима: тежина, висина, густина петне кости и узорак крви. Густина кости била је већа код гојазних у поређењу са децом нормалне тежине и показала је позитивну корелацију са стандардизованим "индексом телесне масе". Адипонектин је показао негативну корелацију са коштаном густином. ОЦ се није разликово у групама. У студији II укључено је 62 гојазне и 46 деце са нормалном тежином као контролна група (узраст 2-9 година стари), изабрана из ИДЕФИЦС студије у Шведској. Ниво укупног ОЦ и уцОЦ се разликово између група. Ниво карбоксилираног ОЦ је био нижи код гојазних. Укупни ОЦ и цОЦ је показао негативну корелацију са ХбА1ц. У студији III испитивало се 22 болесника са тешком АН током 12 недеља интензивног режима исхране за повећане телесне тежине. Тежина, висина, узорак крви и мерење густине костију испитивани су на почетку и крају студије. Ниво укупног ОЦ, уцОЦ и цОЦ је био повећан, али ниједан од облика ОЦ није показао корелацију са променом тежине или нивоом инсулина. Ниво минералног коштаног садржаја је био повишен. У студији IV испитивано је 30 гојазне деце (узраст 7-17 година), од којих су 14 била контролна група, у проспективној, рандомизираној, контролираној студији. Група са 16 деце је била подвргнута тренингу са вибрацијама у току 12 недеља. Мерени параметри су: тежина, висина и узорци крви, густина петне кости, мишиц́на снага и равнотежа. Густина петне кости и тежина је била непромењена, док је равнотежа била побољшана.

Као закључак, ове студије нису могле да потврде позитивни метаболички ефекат ОЦ на

гојазност и инсулинску резистенцију. Запажен је нижи ниво остеокалцина код гојазне

деце.

This thesis is based on the following studies, referred to in the text by their Roman numerals.

I. Relation between bone mineral density, biological markers and anthropometric measures in 4-year-old children: a pilot study within the IDEFICS study

Tubić B, Magnusson P, Swolin-Eide D, Mårild S; IDEFICS Consortium. International Journal of Obesity (London) 2011; 35(Suppl 1): S119-124.

II. Different osteocalcin forms, markers of metabolic syndrome and anthropometric measures in children within the

IDEFICS cohort

Tubic B, Magnusson P, Mårild S, Leu M, Schwetz V, Sioen I, Herrmann D, Obermayer-Pietsch B, Lissner L, Swolin- Eide D; IDEFICS consortium. Bone 2016; 84: 230-236.

III. Increased bone mineral content during rapid weight gain therapy in anorexia nervosa

Tubić B, Pettersson C, Svedlund A, Bertéus Forslund H, Magnusson P, Swolin-Eide D. E-published. Hormone and Metabolic Research, DOI: 10.1055/s-0042-115304.

IV. Whole body vibration intervention: a randomized, prospective, controlled study in children with obesity Tubić B, Zeijlon R, Wennergren G, Mårild S, Dahlgren J, Magnusson P, Swolin-Eide D. Manuscript submitted.

All reprints with the permission of the publishers.

I. Description of an intensive nutrition therapy in hospitalized adolescents with anorexia nervosa.

Pettersson C, Tubić B, Svedlund A, Magnusson P, Ellegård L, Swolin-Eide D, Bertéus Forslund H.

Eating Behaviors 2016; 21: 172-178.

II. Vitamin D status in young Swedish women with anorexia nervosa during intensive weight gain therapy.

Svedlund A, Pettersson C, Tubić B, Magnusson P, Swolin- Eide D. European Journal of Nutrition 2016 E pub ahead of print. In press.

III. The IDEFICS validation study on field methods for assessing physical activity and body composition in children: design and data collection.

Bammann K, Sioen I, Huybrechts I, Casajus JA, Vicente- Rodriguez G, Cuthill R, Konstabel K, Tubić B, Wawro N, Rayson M, Westerterp K, Mårild S, Pitsiladis YP, Reilly JJ, Moreno LA, De Henauw S. International Journal of Obesity (London) 2011; 35(Suppl 1): S79-87.

IV. The relationship between paediatric calcaneal quantitative ultrasound measurements and dual energy X-ray

absorptiometry (DXA) and DXA with laser (DXL) as well as body composition.

Sioen I, Goemare S, Ahrens W, De Henauw S, De Vriendt T, Kaufman JM, Ottevaere C, Roggen I, Swolin-Eide D, Tubić B, Vyncke K, Mårild S. International Journal of Obesity (London) 2011; 35(Suppl 1): S125-30.

V. Validation of anthropometry and foot-to-foot bioelectrical resistance against a three-component model to assess total body fat in children: the IDEFICS study.

Bammann K, Huybrechts I, Vicente-Rodriguez G, Easton C,

De Vriendt T, Marild S, Mesana I.M., Peeters M.W., Reilly

J.J., Sioen I, Tubić B, Wawro N, Wells J, Westerterp K,

Pitsiladis Y, Moreno LA. International Journal of Obesity

(London) 2013; 37(4): 520-6.

A BSTRACT

S UMMARY IN SWEDISH - S AMMANFATTNING PÅ SVENSKA

S UMMARY IN SERBIAN - N AUČNI REZIME

S UMMARY IN SERBIAN - Н AUЧHИ Р EЗИME

L IST OF PAPERS

A BBREVIATIONS ... V

1 I NTRODUCTION ... 1

1.1 Childhood obesity ... 1

1.2 The IDEFICS study ... 1

1.3 Anorexia nervosa ... 2

1.4 Bone tissue ... 3

1.5 The composition of bone ... 3

1.6 Bone cells ... 4

1.6.1 Osteoblasts ... 4

1.6.2 Osteoclasts ... 5

1.6.3 Osteocytes ... 5

1.7 Bone remodelling ... 5

1.8 Bone growth and peak bone mass ... 6

1.9 Bone turnover markers ... 7

1.9.1 Bone formation markers ... 7

1.9.2 Bone resorption markers ... 8

1.10 Osteocalcin ... 8

1.11 The endocrine role of bone tissue, with focus on energy metabolism 10 1.12 Adiponectin ... 14

1.13 Bone mass measurements and body composition ... 14

1.13.1 DXA ... 15

1.13.2 pQCT ... 15

1.13.3 DXL Calscan ... 17

1.15 WBV training ... 19

2 A IMS AND HYPOTHESES ... 22

3 P ATIENTS AND M ETHODS ... 23

3.1 Study Subjects ... 23

3.2 Study designs ... 24

3.3 Osteocalcin – analytical methods ... 25

3.4 Bone measurements ... 29

3.5 WBV training ... 30

3.6 Leonardo mechanography ... 30

3.7 Questionnaires ... 32

3.8 Statistical analysis ... 32

3.9 Ethics ... 34

4 S UMMARY OF PAPERS / R ESULTS ... 35

5 D ISCUSSION ... 41

5.1 Strengths and limitations of the thesis ... 48

6 C ONCLUSIONS ... 50

7 F UTURE PERSPECTIVES ... 51

A CKNOWLEDGEMENTS ... 53

R EFERENCES ... 56

P APERS I-IV ... 73

ALP alkaline phosphatase AN anorexia nervosa

BMAD bone mineral apparent density BMC bone mineral content

BMD bone mineral density BMI body mass index BTM bone turnover marker cOC carboxylated osteocalcin

CTX cross-linked carboxy-terminal telopeptide of type I collagen CV coefficient of variation

DPA dual-photon absorptiometry DXA dual-energy X-ray absorptiometry

DXL dual-energy X-ray absorptiometry and laser

HA hydroxyapatite

HOMA homeostatic model assessment

HR-QCT high-resolution quantitative computed tomography

IDEFICS Identification and prevention of Dietary- and lifestyle- induced health EFfects In Children and infantS

IGF-I insulin-like growth factor -I IR insulin resistance

NTX cross-linked amino-terminal telopeptide of type I collagen

PDGF platelet-derived growth factor

PICP type I procollagen carboxy-terminal propeptide PINP type I procollagen intact amino-terminal propeptide pQCT peripheral quantitative computed tomography SDS standard deviation score

SPA single-photon absorptiometry SSI strength strain index

ucOC undercarboxylated osteocalcin WBV whole body vibration

In the text, bone mass measurements will be used as an umbrella term for

BMD, BMC and BMAD

1 INTRODUCTION

1.1 Childhood obesity

The prevalence of obesity in the world has more than doubled since 1980, with as many as 39% of adults classified as overweight and 13% classified as obese. Worldwide, 42 million children were overweight or obese in 2013 [1].

Never before have so many children and adolescents been obese, although the increase in prevalence seems to be levelling off worldwide [2] and stabilizing in Sweden[3]. Obesity is a disease in itself but it is also a key risk factor for other non-communicable diseases (NCD) such as cardiovascular disease, type 2 diabetes, musculoskeletal disorders and dental disease. In 2001, the proportion of global burden of disease that was attributed to NCDs was 46%; this proportion is expected to increase to 57% by 2020 and to appear in considerably younger age groups [4].

Overweight and obesity are defined as “abnormal or excessive fat accumulation that may impair health” [1]. There is also an arbitrary definition: for adults, overweight is defined as body mass index (BMI) above 25 kg/m ² and obesity is defined as BMI above 30 kg/m ² [1]. While age, sex and genetic susceptibility are non-modifiable, many of the risk factors associated with age and sex can be modified, for example, behavioural factors (diet, physical activity), biological factors (overweight, hyperinsulinaemia) and social (socioeconomic, cultural) factors.

In our society, obese people are often blamed for being irresponsible, lazy and/or undisciplined. But there are also important environmental factors to take into account. There is evidence that when people move to a new environment they may gain in weight [5], for example, moving from one country to another country where obesity is more prevalent. When dealing with obesity, it is important to focus not only on what individuals can do to improve their situation, but also on the whole environment around the individual.

1.2 The IDEFICS study

IDEFICS (Identification and prevention of Dietary and lifestyle-induced

health EFfects In Children and infantS) started in 2006 as a European

multicentre study. The background to the IDEFICS study was the changed

environment for children in Europe with unhealthy dietary habits and a

sedentary lifestyle [6]. The study was in accordance with the WHO strategic directions and recommendations for policy and research about diet, nutrition and prevention of chronic diseases, proposed in 2002 [4]. The IDEFICS study started in 2006 and finished in 2011. The main aim of the study was to investigate the aetiology of diet and lifestyle-related diseases in a large prospective cohort, focusing on overweight and obesity; the secondary aim was to develop, implement and evaluate a community-oriented, population- based intervention programme for primary prevention of obesity in a case- control design setting [7].

At baseline, the prospective cohort comprised of 16 224 children aged two to nine years in eight European countries (Belgium, Estonia, Cyprus, Germany, Sweden, Hungary, Italy and Spain). The Swedish part of the study at baseline consisted of 1 837 children in the Gothenburg area. The study was planned during 2006–2007; thereafter the baseline survey (T0) was performed in 2007-2008. The intervention was executed in 2007–2008 as a case-control design. The follow-up survey (T1) was performed in 2009–2010. During 2010-2011, all the collected data were structured and cleaned before being used in further research. The baseline and follow-up surveys were designed to assess overweight and obesity (using anthropometric measurements and lifestyle questionnaires), musculoskeletal disorders (using qualitative ultrasound examination of the calcaneus) and insulin resistance (using biochemical markers) [7].

1.3 Anorexia nervosa

Anorexia nervosa (AN) is a psychiatric disorder with severe consequences, primarily seen in adolescent girls. The definition of AN according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) criteria [8]

is: significantly low body weight, intense fear of gaining weight, altered body image and lack of recognition of the seriousness of the low body weight. The prevalence of AN is approximately 1% among 17-year-old Swedish girls [9].

The average prevalence rate in the world is approximately 0.3%. Incidence

rates worldwide are up to 8 per 100 000 people per year, though it must be

underlined that even the most well-designed studies underestimate the true

incidence [10]. Approximately 10% of AN patients are men [10]. Most

patients recover from severe AN but there is a relatively high prevalence of

mortality in the group, with a mortality rate of 3% to 18%, depending on the

study [11]. AN leads to severe complications, for example, an increased risk

of suicide, neuropsychiatric co-morbidities, amenorrhea, hormonal imbalance

with potentially lower bone quality and osteoporosis [12, 13].

1.4 Bone tissue

The study of human bone tissue has been of interest since at least the second century of the common era, when Galen of Pergamon (129 CE - 200/216 CE studied the human skeleton in Alexandria, Egypt, and recommended his students to seize every opportunity to do the same [14]. The human skeleton has three distinct functions: firstly, it gives us an upright posture and it is a point of attachment for our muscles, enabling us to move; secondly, the skeleton protects our vital organs, including the brain, heart and lungs; thirdly it serves as a reservoir for calcium, phosphate, lipids and bone marrow. The latest proposed role is a possible endocrine function as a regulator of whole- body glucose metabolism and male fertility [15, 16].

A fascinating characteristic of bone is its unique combination of strength and flexibility. It allows us to walk and at the same time it can tolerate bending, compression and torsion without breaking. To maintain these characteristics, bone is constantly adapting to the physiological and mechanical strains that challenge it.

1.5 The composition of bone

The human skeleton consists of 206 bones, which can be divided into long bones (such as the tibia and femur) and flat bones (such as the cranium and pelvis). The skeleton as an organ consists of cartilaginous joints, calcified cartilage, the marrow space and the mineralized structures including bones.

Mineralized bone is made up of cells, vessels and crystals of calcium compounds (hydroxyapatite). Macroscopically there are two kinds of bone tissues: cancellous (porous/trabecular) and cortical (compact) (Fig 1).

Cancellous bone is found in the metaphysis of long bones and in the

vertebrae; cortical bone is the hard bone located on the surface. The

proportion of cancellous and cortical bone determines the mechanical quality

of bone and its resistance to fractures [17]. Bone mass consists of

approximately 80% cortical bone and 20% cancellous bone, but the

cancellous bone has a much larger area (90% of its total area) exposed to

other tissues and bone marrow. This makes the cancellous bone the main

target for bone mineral metabolism [17].

Figure 1. Human femur cervical neck. Cancellous and compact (cortical) bone.

© https://courses.stu.qmul.ac.uk/SMD/Kb/microanatomy/bone/index.htm

1.6 Bone cells

There are three types of bone cells: osteoblasts, which are responsible for bone formation, osteocytes, which are osteoblasts that are trapped in osteoid, and osteoclasts, which are responsible for bone resorption [18].

1.6.1 Osteoblasts

Osteoblasts produce bone matrix. They originate from multipotent

mesenchymal stem cells, which can differentiate into osteoblasts,

chondrocytes, myoblasts, adipocytes or fibroblasts [19]. As the period for

matrix production ends, 15% of the osteoblasts are entrapped in the newly

produced matrix and become osteocytes. Osteocalcin (OC) is specifically

expressed by osteoblasts [16]. Osteoblasts have receptors for, and produce,

growth factors such as insulin-like growth factors (IGFs) and platelet-derived

growth factor (PDGF). Activity of the osteoblasts is regulated in a paracrine

and autocrine manner by these growth factors. Osteoblasts also have

receptors for hormones such as parathyroid hormone, thyroid hormone,

insulin and growth hormone [18].

1.6.2 Osteoclasts

Osteoclasts are large multinucleated cells derived from hematopoietic cells of mononuclear lineage. The osteoclast is responsible for bone resorption.

During the process of bone resorption, the osteoclasts adhere to the bone matrix and move along the bone surface. Osteoclasts resorb bone by enzymatic proteolysis of bone matrix and acidification of hydroxyapatite (HA) crystals, which are found within the sealing zone [18]. Studies have demonstrated receptors on osteoblasts for thyroid hormone, androgens, calcitonin, insulin, IGF-I and PDGF, among other hormones and growth factors [18].

1.6.3 Osteocytes

Osteocytes are differentiated osteoblasts that become surrounded by bone matrix during bone formation. The exact function of osteocytes is not fully understood; however, it has been suggested that they attract osteoclasts to sites where bone remodelling is required, as a response to bone tissue strain or microdamage [20]. Sclerostin, a glycoprotein, has recently been proposed to be a product of osteocytes acting as an inhibitor of bone formation [21].

Sclerostin levels increase with age [22] and sclerostin is reported to be positively associated with BMD [23]. Mechanical loading has been reported to bear an inverse relation to circulating sclerostin levels [24, 25], whereas immobilization shows a positive correlation [26]. Studies have also demonstrated both positive [27] and non-existing [23] correlation between sclerostin levels and fracture risk.

1.7 Bone remodelling

The process of bone remodelling is vital, as the bone is a dynamic organ with

continuous defects and microfractures which are repaired by the actions of

osteoblasts and osteoclasts. Harold Frost was the first to demonstrate this

process [28]. During remodelling, there is a close interplay between

osteoclasts and osteoblasts, in which they form a basic multicellular unit

(BMU). Between 2% and 5% of cortical bone is remodelled per year. As

cancellous bone has a much larger surface-to-volume ratio, it is more actively

remodelled than cortical bone. The remodelling cycle can be divided into

different phases: resorption, reversal and formation (Fig 2). The resorption

phase lasts for two weeks and is the period when osteoclasts break down

tissue and form a resorption cavity; the reversal phase lasts for four to five

weeks, followed by the formation phase, in which osteoblasts fill the cavity,

which takes approximately three to six months [18, 29, 30].

Figure 2. Bone remodelling. (D.L Kasper, A. S. Fauci, S. L. Hauser, D. L. Longo, J. L. Jameson, J. Loscalzo. Harrison´s principles of internal medicine, 19th edition.) © McGraw-Hill Education. All rights reserved.

1.8 Bone growth and peak bone mass

Peak bone mass can be defined in several ways. At the individual level, the concept of peak bone mass can be defined as the maximum amount of bone accrued during young adulthood. At the population level, peak bone mass is attained when bone outcome has reached a plateau or maximum value, or when age-related changes in bone outcome are no longer positive [31, 32]

(Fig 3). The age at which different bones reach peak bone mass varies

depending on anatomical region [33, 34]. Studies have reported that peak

bone mass in the femur is reached at approximately 16–18 years of age in

girls [33, 35] while the distal radius reaches peak bone mass at 40 years of

age and above in women [34, 35]. There is an uncertainty whether bone mass

is positively or negatively affected by fat mass [36]. There are studies where

the majority of children in the fracture group are overweight [37, 38]. It is

important to understand the interaction between fat and bone when

developing future health strategies to optimise peak bone mass and prevent

fractures [36].

Figure 3. Peak bone mass. Copyright © 1990 Compston JE[39]. The red line represents women and the blue line represents men.

1.9 Bone turnover markers

Bone metabolism mirrors the complex interplay of osteoblasts, osteoclasts and osteocytes. To analyse this interplay, serum and urine assays have been developed. In these assays, biochemical markers are analysed, reflecting the activity of osteoblasts and osteoclasts and the breakdown products of bone tissue. Bone turnover markers (BTM) are divided into markers of bone formation and bone resorption, and are not necessarily tissue- or site-specific (Table 1). The clinical application of these assays is to investigate diseases that alter bone metabolism, for example osteoporosis, to estimate fracture risk, and compliance in order to evaluate pharmaceuticals against the disease in question [40]. The change in levels of BTMs is faster (response within weeks) than the change in BMD that can be measured with, for example, dual-energy X-ray absorptiometry (DXA) or peripheral quantitative computed tomography (pQCT).

1.9.1 Bone formation markers

Bone formation markers can be measured in serum and reflect the different phases of osteogenesis, but they are not all specifically produced by osteoblasts. Osteoid is produced in the early phase of bone formation and 90% of osteoid is type I collagen, which is expressed by osteoblasts. When type I collagen is produced, carboxy-terminal (PICP) and amino-terminal (PINP) propeptides are cleaved of type I procollagen. The circulating levels of these propeptides reflect the amount of newly synthesized type I collagen.

Attainment of peak bone mass

Consolidation Age-related bone loss

Fracture threshold Menopause

Women Men

Bo ne m as s

0 10 30 30 40 50 60

(years(years) Age (years)

Type I collagen is not specific to bone tissue, as it can also be found in tendons and in skin, although it mainly originates from bone. PINP has been chosen by the International Osteoporosis Foundation as the reference marker for bone formation, due to its stability, assay performance and response to treatment [29]. Another bone formation marker is OC, which is produced by osteoblasts and is bound to HA in the mineralized matrix of bone. A third bone formation marker is bone alkaline phosphatase (ALP), which contributes to bone mineralization [29].

1.9.2 Bone resorption markers

Markers of bone turnover are products of the degradation of type I collagen, non-collagenous proteins and enzymes, which are produced during osteoclast activity. When bone is resorbed, peptide fragments of type I collagen, such as carboxy-terminal and amino-terminal cross-linked telopeptides of type I collagen (CTX and NTX, respectively), are released into the blood. Both can be analysed with automated measurement. The International Osteoporosis Foundation has recommended using CTX as the reference marker for bone resorption [29]. During osteoclast activity, tartrate-resistant acid phosphatase type 5 (TRACP5b) is produced by osteoclasts and reflects the number of osteoclasts [29].

Table 1. Bone turnover markers.

1.10 Osteocalcin

Osteocalcin was first isolated by Hauschka et al. [41] in 1975, and confirmed by Price et al. in 1976 [42]. OC was characterized as a gamma- carboxyglutamic acid-containing (Gla) protein from bone; it was found to bind strongly to HA and was confirmed as an important factor in the bone extracellular matrix. Osteocalcin is also known as bone Gla protein (BGLAP) and is the most abundant osteoblast-specific non-collagenous protein [43].

OC is a small protein of 49 amino acids in humans and 46 amino acids in mice. It is synthesized in the osteoblast as a pre-promolecule (Fig 4).

Bone formation markers Bone resorption markers

Type I procollagen carboxy-terminal propeptide (PICP)

Type I procollagen intact amino-terminal propeptide (PINP)

Telopeptides of type I collagen (C-terminal: CTX; N-terminal: NTX)

Osteocalcin (OC) Tartrate-resistant acid phosphatase

(TRACP5b)

Alkaline phosphatase (Total ALP and bone ALP)

Post-translational vitamin K-dependent gamma carboxylation occurs where three glutamic acid (Glu) residues (GLU13, GLU17 and GLU20) are carboxylated into Gla residues by gamma carboxylase. This gamma carboxylation results in a greater affinity for calcium and HA. Eventually, pro-OC is cleaved into carboxylated and undercarboxylated OC. The gamma carboxylation of the three Glu residues is important for the structure and function of the fully cOC, enabling it to bind HA with high affinity and to regulate bone mineral maturation. After the intracellular cleavage, OC is secreted from the osteoblast cell [43]. OC exists in serum in fully carboxylated, partially carboxylated and undercarboxylated forms [44-46].

Carboxylated OC is deposited into the bone extracellular matrix together with calcium and partly released into the circulation, whereas ucOC (where 0-2 Glu residues are carboxylated) is mainly released into the circulation [47]. In 1984, Brown et al. demonstrated evidence of OC acting as a bone formation marker, which could be used to evaluate the treatment of postmenopausal osteoporosis [48].

Figure 4. Synthesis of OC in osteoblasts. Vitamin D stimulates the transcription of the BGLAP gene. The preproosteocalcin is cleaved to proosteocalcin. A

proportion of proosteocalcin is then carboxylated in a vitamin K-dependent

process. The final products are carboxylated and undercarboxylated OC that are

released from osteoblasts in a calcium-dependent process. Carboxylated OC is

mainly involved in the mineralization of bone matrix while undercarboxylated OC

is released into the circulation. VDR=vitamin D receptor, VDRE=vitamin D

responsive element, Gla=gamma-carboxyglutamic acid-containing protein,

Glu=undercarboxylated OC. Copyright © 2013 Aurora Patti et al.[49].

1.11 The endocrine role of bone tissue, with focus on energy metabolism

Until recently, the endocrine role of bone has been regarded as a regulator for calcium- phosphorus homeostasis and as a source of hematopoietic cells [15].

During the last 15 years, scientists have studied and proposed that bone acts as an endocrine organ in a wider context.

This research field was initiated by Gerard Karsenty and his group with a publication in 1996 in which Ducy et al. [50] generated an OC-deficient mouse model and demonstrated that the mice developed a phenotype marked by higher bone mass and bone with improved quality. Absence of OC led to increased bone formation and OC was proposed as a negative regulator of bone formation. In 2000, Ducy et al. [51] proposed a common regulation of bone mass, body weight and gonadal function. They studied leptin-deficient and leptin-receptor-deficient mice that were both obese and hypogonadal, and they demonstrated that both mouse models had increased bone formation.

Major progress was made by Lee et al. [52] when they demonstrated that mice null for Esp (also known as Ptprv, a gene encoding the osteotesticular protein tyrosine phosphatase, OST-PTP) were hypoglycaemic and protected from obesity and diabetes because of an increase in β-cell proliferation, insulin secretion and insulin sensitivity (Fig 5). Their phenotype was fully corrected by crossing Esp-null mice with OC heterozygous mice. Further results from the same study presented in vitro experiments where OC- producing osteoblasts enhanced insulin production by pancreatic islets, insulin sensitivity and adiponectin expression in adipocytes. In contrast, OC ^–/– knockout mice, which did not have any OC production, presented with obesity and glucose intolerance. The positive effects were attributed to ucOC.

Carboxylated OC and ucOC are proposed to have distinct functions. The two OC forms have different negative charges and calcium-binding properties, which could explain their diverse biological functions [47]. Ferron et al. [53]

reported in 2008 that OC in vitro induces expression of insulin genes, increase β-cell proliferation and induces adiponectin expression (Table 2).

They could also demonstrate decreased OC levels, increased insulin levels

and decreased fat pad mass in mice receiving OC through an implanted

pump. In 2010 Ferron et al. [54] conducted further studies of OC, in which

they demonstrated that insulin signalling in osteoblasts enhances OC activity

through a promotion of bone resorption in osteoblasts leading to an increased

decarboxylation of OC resulting in increased metabolic activity.

Karsenty´s research group expanded its studies of OC based on the hypothesis that bone mass, body weight and gonadal function are regulated by common pathways [51]. Oury et al. [55] established in 2011 that ucOC favours male fertility by promoting testosterone synthesis by Leydig cells.

Otani et al. [56] demonstrated both in vitro and in vivo (in mice) in 2015 that ucOC stimulates adipocytes to produce adiponectin. Oury et al. [57] also suggested that, in mice, OC influences brain development and function by preventing anxiety and depression as well as favouring learning and memory.

These effects are the result of OC suppression of GABA biosynthesis and favour the expression of serotonin and catecholamine synthesis.

When trying to elucidate the function of a hormone, it is of great importance to demonstrate which receptor it binds to. In the case of OC, the receptor is proposed to be a G-protein coupled receptor: GPRC6A [58], although the structure and function of the receptor has not yet been reported [47]. There have been some studies that contribute to the theory that GPRC6A is a potential OC receptor [58-60], but other groups have not demonstrated evidence in favour of this theory [61, 62]. If GPRC6A is to be confirmed as the OC receptor in the future, then most certainly more functions will appear for OC, because GPRC6A is expressed widely [63]. So far, no other receptor has been identified as a potential receptor for OC.

Figure 5. Different functions of OC. It is suggested that undercarboxylated OC

stimulates β-proliferation and insulin secretion in the pancreas, energy expenditure in

muscle, and insulin sensitivity in muscle, liver and adipose tissue. Through the

activation of the receptor GPRC6A it also promotes testosterone synthesis in Leydig

cells in the testis, leading to increased male fertility. Copyright © 2012 Karsenty et al.

[65] [66]

[67] 2011[68] [69]

[70] -Berry et al., 2012[71]

[72] [52]

[53] [65]

[67] [73]

014[74] [75] -Berry et al., 2012[71]

[72] [53]

[65] [66]

[67] [76] [77] 15[70]

[74] T abl e 2. A ss oc ia ti ons be tw ee n bone turnove r m arke rs a nd gl uc os e m et abol is m a nd adi pos it y i n c hi ldre n a nd a dul ts . as u re o f gl u cos e ta b ol is m an d po si ty

N T ot al OC uc O C cO C Ot h er Me as u re s of b on e res orp ti on Re fe re n ce A -IR 2493 M + F in ve rs e

-

- - - Sa le em e t a l., 2 01 0[6 5] 1597 M inve rs e in ve rs e - - no as soc . (T R A C P ) Ik i e t a l., 2 01 2[6 6] 1010 M in ve rs e - - - - Ki nd bl om e t a l., 2 00 9[6 7] 580 M + F in ve rs e - - - - Gr av en st ei n et a l., 2 01 1[6 8] 348 M + F no as soc ia ti on in ve rs e - - Sh ea e t a l., 2 00 9[6 9] 141 M ( chi ldr en) no as soc ia ti on - - - - Jü ri m äe e t a l., 2 01 5[7 0] 106 M + F ( chi ldr en) no as soc ia ti on inve rs e - - - Bo uc he r- Be rr y et a l., 2 01 2[7 1] 36 F ( chi ldr en) no as soc ia ti on - - pos it ive ( P IC P ) pos it ive ( N T X ) Mi sr a et a l., 2 00 7[7 2] su li n se ns it iv it y ES P -/E S P - mo us e pos it ive - Le e et a l., 2 00 7[5 2] -cel l p ro li fer at io n mo us e pos it ive - Fe rr on e t a l., 2 00 8[5 3] su li n (f as ti ng ) 2493 M + F in ve rs e - - - - Sa le em e t a l., 2 01 0[6 5] 1010 M in ve rs e pos it ive (P IN P ) Ki nd bl om e t a l., 2 00 9[6 7] 380 M + F in ve rs e - - - no as soc . (N T X ) Pi tt as et al ., 2009 [7 3] 68 M + F no as soc ia ti on - no as soc ia ti on - no as soc . ( P IN P ) Vi lj ak ai ne n et a l., 2 014 [7 4] 140 M + F ( chi ldr en) no as soc ia ti on no as soc ia ti on in ve rs e - - Po ll oc k et a l., 2 01 1[7 5] 106 M + F ( chi ldr en) no as soc ia ti on no as soc ia ti on - - - Bo uc he r- Be rr y et a l., 2 01 2[7 1] 36 F ( chi ldr en) no as soc ia ti on - - pos it ive ( P IC P ) pos it ive ( N T X ) Mi sr a et a l., 2 00 7[7 2] su li n mo us e pos it ive - Fe rr on e t a l., 2 00 8[5 3] uc os e (f as ti ng ) 2493 M + F in ve rs e - - - - Sa le em e t a l., 2 01 0[6 5] 1597 M in ve rs e in ve rs e - - no as soc . (T R A C P ) Ik i e t a l., 2 01 2[6 6] 101 0 M in ve rs e - - - no as soc . (P IN P ) Ki nd bl om e t a l., 2 00 9[6 7] 328 M + F in ve rs e - - - - Ka na za wa e t a l., 20 09 [7 6] 290 M + F no as soc ia ti on - - - - Bu da y et a l., 2 01 3[7 7] 141 M ( chi ldr en) no as soc ia ti on - - - - Jü ri m äe e t a l., 2 015 [7 0] 68 M + F no as soc ia ti on - no asso ci at io n - no as soc . ( P IN P ) Vi lj ak ai ne n et a l., 2 01 4[7 4]

[75] -Berry et al., 2012[71]

[78] [65] [69]

[76] [76] [70] -Puig et al., 2010[79]

[80] [76] -Berry et al., 2012[71]

[81] [65] [73]

[70] -Berry et al., 2012[71] -Puig et al., 2010[79] esta et al., 2010[82]

[53]

as u re o f gl u co se ta b ol is m an d po si ty

N To ta l O C uc O C cO C Ot h er Me as u re s of b on e res orp ti on Re fe re n ce 140 M + F ( chi ldr en) no as soc ia ti on no as soc ia tio n no as soc ia ti on - - Po ll oc k et a l., 2 01 1[7 5] 106 M + F ( chi ldr en) pos it ive no as soc ia ti on - - - Bo uc he r- Be rr y et a l., 2 01 2[7 1] 64 M + F ( obe se ) in ve rs e - - no as soc . (P INP ) - Ig le si as e t a l., 2 01 1[7 8] ip on ec ti n 2493 M + F pos it ive - - - - Sa le em e t a l., 20 10 [6 5] 348 M + F pos it ive no as soc ia ti on pos it ive - - Sh ea e t a l., 2 00 9[6 9] 149 F pos it ive - - - - Ka na za wa e t a l., 20 09 [7 6] 179 M no as soc ia ti on - - - - Ka na za wa e t a l., 20 09 [7 6] 141 M (c hi ld re n) no as soc ia ti on - - - - Jü ri m äe e t a l., 2 01 5[7 0] 103 M + F - - in ve rs e - - Pr at s- Pu ig et a l., 2010 [7 9] B od y fat 443 M + F in ve rs e ( in F ) - - - - Sh ea e t a l., 2 01 0[8 0] 179 M in ve rs e - - - - Ka na za wa e t a l., 20 09 [7 6] 106 M + F ( chi ldr en) in ve rs e no as soc ia ti on - - - Bo uc he r- Be rr y et a l., 2 01 2[7 1] 79 M + F ( chi ldr en) in ve rs e - - - - Wa ng e t a l., 2 01 4[8 1] I 2493 M + F in ve rs e - - - Sa le em e t a l., 2 01 0[6 5] 380 M + F in ve rs e - - - no as soc . (N T X ) Pi tt as et al ., 2009 [7 3] 141 M (c hi ld re n) In ve rs e - - - - Jü ri m äe e t a l., 2 01 5[7 0] 106 M + F ( chi ldr en) in ve rs e in ve rs e ( in M ) - - - Bo uc he r- Be rr y et a l., 2012 [7 1] 103 M + F ( chi ldr en) - - pos it ive - - Pr at s- Pu ig e t a l., 2 01 0[7 9] 83 M - no as soc ia ti on no as soc ia ti on no as soc ia ti on - Fo re st a et a l., 2 01 0[8 2] t m as s mo us e in ve rs e - Fe rr on e t a l., 2 00 8[5 3] Mo di fi ed f ro m B oo th e t al ., 2013, N at ur e R ev En do cr in ol .[8 3] OC = os te oc al ci n, uc O C = unde rc ar boxyl at ed os te oc al ci n, cO C = car bo xy lat ed O st eo cal ci n, HOM A -IR = hom eos ta ti c m ode l a ss es sm ent - in su lin r es is ta nc e, F = fe m al e, M = ma le .

On ly s ta ti st ic al ly s ig ni fi ca nt ( P < 0. 05 ) as so ci at io ns ( po si ti ve o r in ve rs e) a re s ho wn .

No m ea su re m en ts we re r ep or te d.

1.12 Adiponectin

Adiponectin was first discovered by Scherer et al. [83] in 1995. The first name given was Acrp30, (adipocyte related complement protein of 30 kDa) which was later renamed to adiponectin. Adiponectin is a protein made only by adipocytes and its secretion is enhanced by insulin. Even in the first publication, adiponectin (Acrp30) was proposed as a possible factor in energy homeostasis as well as in fat and glucose metabolism [83]. Since then, more than 15 000 papers containing adiponectin in the title or abstract have been published. Adiponectin has several important functions but the primary ones are its insulin-sensitizing qualities in peripheral tissue and its protective effects against inflammation and apoptosis [84]. A number of studies have demonstrated an inverse association between adiponectin and insulin levels, glucose concentration and obesity [85-87]. Yamauchi et al. [88] have demonstrated in mice studies that administration of adiponectin decreases insulin resistance and Gao et al. [89] have suggested a causal relationship between adiponectin and insulin resistance. The endocrine role of OC has been described in numerous publications in the same context as adiponectin, where the two proteins have been proposed to interact. OC and its positive metabolic actions have been suggested to be partly mediated through an induction of adiponectin production in adipocytes and a decrease in adipocyte size [52, 56].

1.13 Bone mass measurements and body composition

In 1895, Wilhelm Conrad Röntgen discovered what he called “X-rays” which paved the way for the first X-ray machines. In 1901 he received the first Nobel prize in physics. With the invention of the radiographic technique it became possible to visualize different qualities of bone, for example osteoporosis. In 1959, a workshop was held on bone densitometry in Bethesda, USA. The organizing researchers conducted an overview of all existing literature, which was published in 1962 by Garn [90]. They were surprised by the volume of research that had been conducted (125 papers during the period 1897–1961). In 1963, Cameron and Sorensen [91]

described the single-photon absorptiometry (SPA) method, which was also

described by Nilsson in 1966 [92]. This method was peripheral and enabled

calculation of the mineral content in grams of calcium hydroxyapatite per

centimetre length of bone. Later on, dual-photon absorptiometry (DPA) was

developed from the SPA method. DPA uses two photon sources and makes it

possible to measure the hip and spine, because they are located centrally in the body. Subsequently, DPA was replaced with DXA.

1.13.1 DXA

Dual-energy X-ray absorptiometry (DXA) was first described by Cullum et al. [93] in 1987 and is now the gold standard of clinical bone densitometry techniques. One main difference compared to the DPA technique is that the photons are produced from a low-dose X-ray tube instead of from a source of radionuclides. DXA sends out X-rays in two distinct energy levels and is able to measure two different tissue components, bone and soft tissue, where it is assumed that the relationship between lean soft tissue and adipose tissue is constant. DXA is a two-dimensional X-ray method in contrast to three- dimensional methods such as pQCT [94].

DXA scans are used to measure BMD at hip, spine, total body and forearm.

The three major clinical roles of the DXA scan are to diagnose osteopenia, to assess the patient’s risk of fracture and to monitor the effect of treatment [95]. Furthermore DXA is used to analyse body composition such as “fat mass”, “lean mass” or “fat-free soft tissue” [96].

1.13.2 pQCT

Peripheral quantitative computed tomography (pQCT) is primarily used in

research, although the technique is also becoming more popular in the

clinical setting (Fig 6). The advantage of this method is its ability to measure

bone geometry and volumetric bone density (three-dimensional X-ray

method). It is a peripheral method which most commonly measures sites in

the appendicular skeleton, for example, the tibia and radius. Recently, high-

resolution quantitative computed tomography (HR-QCT) has been

introduced, mainly in research. The HR-QCT technique is an improvement

on the QCT technique, primarily because of its higher resolution, allowing

estimation of micro-architectural properties such as cancellous bone size and

number, and because of its lower radiation dose [97].

Figure 6. Patient during a pQCT of tibia (above) and output on computer screen

(below). With permission from the patient.

1.13.3 DXL Calscan

The dual-energy X-ray absorptiometry and laser technique (DXL) Calscan system (Fig 7) is based on the DXA method with the addition that the total width of the heel is also measured with laser and used in the calculation of BMD [98]. With the DXL Calscan system (paediatric version) calcaneal BMD [99], and a new parameter called bone mineral apparent density (BMAD) is measured [100, 101]. The DXL Calscan system predicts osteopenia well and the correlation with the whole-body DXA method is high [99, 102, 103]. The DXL Calscan has been used to diagnose osteoporosis in adults and children[100] and it is used in conjunction with measurements using axial DXA technology[103, 104].

Figure 7. DXL Calscan bone densitometer.

The available methods have their strengths and limitations. A summary of

several of the methods is given in Table 3.

Table 3. Summary and comparison of different bone measurement techniques.

Te ch ni qu e Si te Ex am in at io n tim e ( m in ) Ra di at io n dos e (µ S v) Pr ec is io n (C V % ) Ad va nt ag es Li m it at io ns DXA Lu m ba r sp in e To ta l b od y Pr ox im al fe m ur 3– 10 pe r bodypa rt 0. 4– 4 1– 2 <1 1- 2 0. 2- 5. 4

1. H igh pr ec is ion 2. P ae di at ri c re fe re nc e da ta a va il abl e 3. L ow r adi at ion dos e 4. R api d sc an ti m e 5. C an as se ss body co m po si ti on

1. S iz e- de pe nde nt me as ur eme nt s 2. I nt egr al me as ur eme nt o f can cel lo us an d co rt ical b on e 3. S ens it ive to body com pos it ion ch an ges DXL Ca lc an eu s <5 <0 .2 ( ad ul t) <0 .1 2 (c hi ld re n)

1. 2 2. 4- 9. 8 (a ge s 2- 7) 1. S ee D X A 1 –4 2. P or ta ble m ac hin e 3. S hor t m ea sur ing tim e ( ≈ 2 m in )

1. S ee D X A 1 –2 2. O nl y ca lc ane us can b e m easu red pQ C T Ra di us Ti bi a 5– 15 < 1 .5 – 4 pe r scan < 1 .5 – 4 pe r scan

0. 8– 1. 5 3. 6– 7. 8 (a ge s 3– 5) 1. 3– 1. 8 (a ge 1 2) 1. S iz e- in de pe nd en t 2. M ea sur es c or ti ca l an d can cel lo us bo ne sep ar at el y 3. M ea sur es bone , mu sc le a nd f at

1. S ki ll ed st af f re qu ire d fo r ex am in at io n 2. L ong sc an ti m e 3. C an onl y me as ur e pe ri phe ra l s it es Th e ta bl e is a dap ted f ro m “ B on e D en si to m et ry in G ro w in g P at ien ts: G ui del in es fo r C li ni cal Pr ac ti ce ” (c ha pt er 2 ) an d Sö de rp al m e t a l. [9 8] . DXA= du al -en er gy X -ra y ab so rp ti om et ry , DXL = du al -en er gy X -ra y ab so rp ti om et ry a nd la se r, p Q C T = pe ri phe ra l q ua nt it at ive c om put ed to m og ra ph y.

1.14 Impact of mechanical loading on bone

In the earlier years of skeletal physiology, during the 1960s, it was thought that the effector cells in bone tissue (chondroblasts, fibroblasts, osteoblasts and osteoclasts etc.) alone determined the characteristics of bones, joints, fascia, ligaments and tendons. This was attributed to regulation from non- mechanical agents such as calcium, vitamin D, hormones, genes and age.

This view was changed by Harold Frost and his mechanostat theory, which suggests that all skeletal organs (cancellous bone, cortical bone, growth plate, articular cartilage, tendons, ligaments and muscle) adapt their structure, strength and stiffness to the voluntary loading that is exerted on them. There are three states bone can have: disuse, adapted and overload which finally results in fracture. Between the disuse and adapted state and between adapted and overload there are threshold ranges for strain (bone deformation during loading). During the disuse state there is bone loss; in the adapted state, bone mass is held in steady state; and in the overload state, modelling occurs, that is to say, bone mass is increased [105]. This theory has been examined in intervention studies, with reports of a positive effect of physical activity on bone quality in children [106-108].

1.15 WBV training

The modern form of WBV was a further development of, rhythmic

neuromuscular stimulation, developed by the Soviet scientist Vladimir

Nazarov [109]. In 1960, Bierman [110] conducted a case-control study with

handheld cycloid vibration equipment, in which paravertebral muscles in the

lower back and posterior aspects of lower extremities were stimulated. He

could demonstrate that trunk flexibility increased in the vibration group,

which he attributed to muscle relaxation. In a well-known study by Rubin et

al. [111], the hind limbs of sheep were stimulated with vibration therapy (30

Hz, 20 minutes per day, five times per week for one year) and showed a

twofold increase in bone formation, a 30% increase in cancellous bone

volume and a 34% increase in cancellous bone density. Vibration therapy

with the Galileo 2000 vibration training device (Novotec, Pforzheim,

Germany) (Fig 8) has now been incorporated in the current space

programme, where the aim is to prevent muscle and bone loss in astronauts

[112].

Figure 8. A co-researcher standing in foot position 2 on a Galileo Med Basic WBV machine.

There are two principal WBV techniques in the commercially available devices: the vertical technique, in which vibration is transferred to both feet synchronously, and the side-alternating technique, in which the right foot is synchronously lowest when the left foot is highest, and vice versa [113]. The side-alternating technique is more analogous to walking and running, and may be associated with less risk of negative side effects [114].

During the last decade, WBV training has gained in popularity and nowadays a WBV device can be found in every larger fitness centre. With the ongoing popularity of these devices, an increasing number of studies have been performed and a guideline document on how to conduct and report studies has been published [115].

WBV has been introduced as an alternative or supplement to regular physical

activity and as a treatment option for several clinical conditions associated

with loss of musculoskeletal mass, including osteoporosis and muscle

strength [116-118]. It has also been evaluated as a treatment option for

improving mobility in severely motor-impaired children and adolescents

[104, 119-121]. It has been demonstrated that WBV training has a positive

effect on fasting blood glucose [122-125], insulin [122], HbA1c [123, 126],

weight [127], bone formation and resorption [128-130].

Nonetheless, there is a limited number of studies on WBV training in

children, and specifically in children with obesity. The only study performed

in young children with obesity was by Erceg et al. [131] who investigated the

effects of WBV on bone mass, glucose and bone metabolism in 10-year-old

overweight boys. Erceg et al. [131] could only demonstrate a positive effect

of WBV on BMC and BMD.

2 AIMS AND HYPOTHESES

General aim

• The general aim of the thesis was to investigate the interplay between bone and fat tissue, with focus on OC, through clinical studies in children and adolescents.

Specific aims.

• To clarify the role of OC in relation to weight, with focus on ucOC and cOC, in children and adolescents (papers I–IV).

• To gain insight into how obesity and underweight affect bone and fat tissue in children and adolescents, with focus on ucOC and cOC (papers II–IV).

• To explore the relationship between weight, anthropometric data and bone mass (papers I, III and IV).

• To study the effect of WBV on energy and bone metabolism, anthropometric measurements, muscle parameters and BMD (paper IV).

Hypotheses

• OC, and specifically ucOC, is associated with a favourable metabolic profile in children and adolescents (papers I–IV).

• Weight has a positive effect on BMD (paper I and paper III).

• Adiponectin is inversely associated with bone mass in children and adolescents (paper I and paper III).

• WBV training increases muscle strength, which in turn

improves the metabolic profile and BMD (paper IV).

3 PATIENTS AND METHODS

3.1 Study Subjects

Paper I

Subjects in paper I were included from the Gothenburg cohort (1825 children between two and nine years of age) in the IDEFICS study and were examined during the baseline survey (n=1825). All children in the age range 4.0 to 5.0 years were selected. This gave a total of 41children (28 boys) in this sub- study. Each of these 41 subjects underwent a heel DXL scan. Twenty-nine subjects also had blood samples taken during the IDEFICS examinations.

The group was divided into normal-weight and overweight/obese according to BMI standard deviation score (BMI-SDS) [132]. Because the IDEFICS examinations were performed several months before the heel DXL, we also measured height and weight at the time of the heel DXL examination. The correlation between these two measurements was r=0.79 (P=0.001).

Paper II

In paper II the subjects were also included from the Gothenburg cohort in the IDEFICS study. The 62 children who were the most overweight or obese and had provided a sufficient blood sample were included, as well as 46 normal weight children. The inclusion criterion was overweight or obesity according to Cole et al. [133]. The children were matched for age (maximum difference of six months), gender and the month when the blood samples were taken.

All included subjects were between two and nine years of age.

Paper III

Twenty-seven participants met the eligibility criteria and were asked to

participate during the inclusion period (January 2012 to July 2014). Inclusion

criteria for the study were age between 16 and 24 years and diagnosis of AN

according to the DSM IV [134]. Exclusion criteria were diabetes mellitus,

inflammatory bowel disease or any physical condition that demanded care at

an internal medicine department. Twenty-five patients agreed to participate

of whom three did not complete the study: two because they did not want to

go through the 12-week programme and one due to reconsideration of the AN

diagnosis. Thus, 22 patients completed the study according to protocol.